超越正确性：通过可扩展的智能体判断标注增强代码大模型的架构推理能力
Beyond Correctness: Enhancing Architectural Reasoning in Code LLMs via Scalable Labeling with Agentic Judgment
AI总结：
针对代码大模型缺乏架构理解的问题，提出智能体判断流水线，利用强LLM作为专家架构评估的代理，通过两个判断器（ACJ和AQJ）实现可扩展标注，微调模型在SWE-bench上提升高达540%，并展现跨语言泛化能力。
摘要：
大语言模型（LLM）已显著改进软件工程，但实际开发需要架构理解。这种理解的人工标注成本过高，且无法仅通过测试验证。我们提出一种智能体判断流水线，使用强LLM作为专家架构评估的可扩展代理，包含两个判断器：架构复杂度判断器（ACJ）评估任务所需的代码库特定架构理解，架构质量判断器（AQJ）通过基于源代码的准则评估补丁对仓库特定架构约定的符合程度。在3360个精选实例上微调Qwen3-8B/14B/32B，在SWE-bench Verified上实现了高达27.2%的解决率——比基础模型提升540%，比未过滤微调提升256%。同时，训练后的模型实现了强大的跨语言泛化能力和架构补丁质量的一致改进。
英文摘要：
LLMs have substantially improved software engineering yet real-world development requires architectural understanding. Such understanding is prohibitively expensive to label manually and impossible to verify through tests alone. We propose an agentic judging pipeline using a strong LLM as a scalable proxy for expert architectural evaluation, comprising two judges: the Architecture Complexity Judge (ACJ), which estimates codebase-specific architectural understanding a task demands, and the Architecture Quality Judge (AQJ), which evaluates patch conformance to repository-specific architectural conventions via source-grounded rubrics. Fine-tuning Qwen3-8B/14B/32B on 3,360 curated instances achieves resolved rates of up to 27.2% on SWE-bench Verified - up to 540% over the base model and 256% over unfiltered fine-tuning. Meanwhile, the trained models achieve strong cross-language generalization and consistent improvements in architectural patch quality.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
